This browser can't run the on-device speech engine.
It needs Web Workers, microphone access, and WebAssembly. Please use an up-to-date
Chrome, Edge, Brave, Firefox, or Safari.
Preparing speech model…
First time only — the AI model downloads once (~tens of MB) and is then cached on your device.
Speak naturally and pause between thoughts — Scribe transcribes each phrase. Punctuation & capitalization are automatic.
🔒 100% on-device. Your voice is transcribed by an AI model running inside this browser — audio and text never leave your computer. Transcript is saved only in this browser.